Rolling Time-domain State Estimation For Uncertain Networked Systems

With the development of communication network and the progress of computer technology,the control mode begins to change from the traditional point-to-point mode to the networked control mode.Then,control systems called Networked Control Systems(NCSs)spring up,which can through a shared network realize the information transmission between the nodes such as controller,sensor and estimator.At present,although the research on NCSs has made a lot of achievements,there are still some problems that have not been effectively solved.For example,how to solve the problem of optimal state estimation under constraints,and how to get the information fusion strategy and consistency algorithm of some sensors in distributed system,and so on.Based on the above,this paper studies some problems in the state estimation of NCSs,such as model error,measurement loss,quantization distortion and multi-sensor information fusion.The main research contents are as follows:1.This paper designs a robust moving horizon estimation method for NCSs with random packet loss and model error.First,under the unified framework of packet loss probability,model error,transmission noise and system constraints,NCSs is modeled as a class of norm-bounded systems with uncertain parameters.Then,a robust moving horizon estimation is designed based on the moving optimization principle and the least square principle.In this method,the stochastic parameters are applied to the solution of the time-varying covariance matrix,which not only ensures the constraint processing ability of the estimator,but also obtains the optimal estimation state of the system.Finally,the performance of the estimator is analyzed,and the effectiveness of the algorithm is verified by simulation.2.This paper designs a robust distributed moving horizon estimation method for distributed NCSs with random packet loss.First,an input holding strategy is used to model the NCSs as an system with uncertain parameters.Then,based on the moving optimization principle and the scalar weighted linear minimum variance strategy,the performance index and measurement data of each sensor node are weighted.And by solving the quadratic optimization problem in a fixed window,a local estimator is obtained.The estimated state obtained by this method has local optimality,which avoids the problem of the system stability when multiple estimated results are weighted directly in the traditional fusion strategy,and improves the estimation efficiency.Finally,the performance of the estimator is analyzed,and the effectiveness of the algorithm is verified by simulation.3.This paper designs a robust distributed moving horizon estimation method for distributed NCSs with random packet loss and quantization.Suppose the system uses a logarithmic quantizer.First,the input holding strategy and sector bounded analysis are used to model the NCSs as a class of norm-bounded systems with uncertain parameters.Then,based on the principle of moving optimization and information fusion strategy,a local estimator is obtained by solving a minimum quadratic optimization problem with random parameters.By using an alternative matrix to replace the error covariance matrix,this method simplifies the calculation process and further improves the estimation efficiency.Finally,the performance of the estimator is analyzed,and the validity of the algorithm is verified by simulation.